Cisco enabling partners for AI Defense release
Between now and March, Cisco will be enabling partners to build the competencies and the capabilities needed for the release of Cisco AI Defense to have better engagement with customers.
Cisco recently unveiled AI Defense to help organizations in their AI deployment. As organizations continue to develop more AI use cases, one of their biggest concerns is the safety and security of using AI. Cisco AI Defense safeguards enterprises against the misuse of AI tools, data leakage and increasing cyberthreats.
While existing security offerings do provide some form of safety and security, most of them are not equipped to handle during the development, deployment and ongoing use of AI apps.
“We want to simplify how organizations would manage the safety and security of their workloads. Often, the reason they’re not able to move workloads into production fast enough is because they’re concerned about safety and security. We’re making it incredibly simple for them to be able to deploy security to protect their use of AI,” said DJ Sampath, VP of Product, AI Software and Platform at Cisco.
[Related: Cisco amplifies AI Security with AI Defense]
According to Raymond Janse van Rensburg, Vice President Networking Sales and Solutions Engineering Asia-Pacific, Japan, and Greater China at Cisco, the traditional application stack had three layers, infrastructure, data, and applications. But now with AI, there’s a fourth layer added to that, which is the model. However, van Rensburg pointed out that it's not just one model as there will be multiple models that will be in multiple clouds.
“Here's the thing about models. By definition, they are nondeterministic. You’re not always going to know what the output of a model will be. And these models will be dynamic, always changing, always evolving, and that will create a new set of risk factors. We must protect the organization and the models against these risk factors from a model as well as an application perspective. Because we all know that when these models break, bad things will happen,” explained van Rensburg.
Organizations currently have two sets of issues they need to deal with, which are safety and security. For safety, van Rensburg explained that models could have issues with toxicity and bias. On security, he said that organizations are now having to defend themselves against threats and attacks on the model from the outside, where the bad actors are changing or trying to change the behavior of the model.
“We want to simplify how enterprises would manage safety and security of their AI applications. A lot of times the reason why they're not able to move something into production fast enough is because they're concerned about safety and security and we're making it incredibly simple for them to be able to deploy security to protect the use of AI,” he added.
How Cisco AI Defense works
“AI Defense builds on the cutting-edge work of Robust Intelligence, an AI security company recently acquired by Cisco. They pioneered several breakthroughs, from supply chain security, algorithmic jailbreaking, and introduced the industry’s first AI Firewall. Combined with world-class detection models powered by Scale AI and integrated telemetry from Cisco Talos, AI Defense offers unmatched protection and delivers unparalleled AI security capabilities,” said Sampath (pictured above).
Sampath explained that the single, end-to-end solution embeds purpose-built AI security technology into the existing network visibility and enforcement points in the Cisco Security Cloud.
“By performing AI security at the network level, we empower security teams to provide consistent and reliable protections for enterprise AI applications across any number of clouds and models. This saves developers valuable time and resources which can be spent pushing AI innovation even further,” he said.
On how AI Defense gets its data to detect and secure AI models, Sampath mentioned that the core technology that is enabling this is basically a technique called the Tree of Attacks with Pruding, which was pioneered by Robust Intelligence. Robust Intelligence is a startup Cisco acquired last year.
Interestingly, Sampath also commented that AI Defense will also be combing this data with the latest threat intelligence they get from Talos as well as Splunk, which Cisco acquired last year.
At the same time, with organizations looking to consolidate and improve their security solutions, Sampath said Cisco is constantly innovating to apply AI to make security operations a lot simpler.
“We announced our XDR AI assistant to be able to help enterprises triage and manage these alerts that are showing up. This reduces the amount of alert fatigue that comes into play. We're also seamlessly integrating AI Defense outputs into Splunk using a technology add on, and that allows enterprises to use Splunk's AI assistant to be able to mitigate any amount of overload that they might have from the alerts that they're seeing so that they can simplify their remediation of response. We also see the signals that we're emitting from the AI Defense of being very high fidelity, so enterprises don't have to necessarily deal with a lot of faults,” explained Sampath.
Deploying Cisco AI Defense
While some components of Cisco AI Defense are already available to some customers for testing, the full availability of the solution is expected to be in March this year. As such, Cisco is now actively preparing its partner ecosystem on how they can best offer this product to customers.
In Asia Pacific, Dave West, President for Asia Pacific, Japan and Greater China (APJC) at Cisco commented that partners will be critical to reaching out to customers with the new solution. West said that Cisco is already working with partners closely across the board and has already started briefing them on AI Defense.
Sampath also mentioned that Cisco has already walked key channel partners through the whole launch announcement, including the training they will need to take, even before the launch. This is to make sure that partners are aware of the product that’s being brought to the market.
“We’re working with partners in the US and across APJC as well. We’ve had briefings with them and close conversations with partners like Data#3 and a bunch of other folks. On the MSSP side of the house, we’re bringing all these solutions baked into the Cisco Security Cloud, and more specifically the Security Cloud Control Console. This will allow MSSPs to be able to provision and manage all of these capabilities. All of Cisco Security capabilities using that console is coming to market around summer,” said Sampath.
Meanwhile, West highlighted that 97% of Cisco’s business in the APJC region flows through the channel ecosystem. He added that Cisco has a lot of very attractive and successful partners and they’re going to continue to grow.
“We just so happen to have a lot of Robust Intelligence partners as well because the founder of Robust Intelligence is Japanese, so there's a lot of affinity to this part of the world. As we combine these capabilities for AI Defense, you'll see us ramp very quickly. March 2025 is sort of that stepping off point where we go, but between now and March, we'll be enabling our partners to build the competencies and the capabilities. This includes ramping up our teams in the next 30 to 45 days so that we’re ready to go in March successfully. We’ve got new fiscal years in Japan and in India as we cross the March to April boundary, so it’ll be a really good time,” explained West.
Cisco is also currently revamping its partner program. The Cisco 360 Program which was announced at the Cisco Partner Summit in November 2024, will be effective in February 2026. Rodney Clark, Global Channel Chief (SVP Partnerships and SMB) at Cisco was also recently in Singapore and Tokyo to meet with partners and Cisco Partner teams to answer questions about the program as well as discuss more opportunities with partners in the region.